A continuous, implanted glucose sensor for diabetics would provide warnings of hypoglycemia and information for improved insulin replacement either by conventional injection, mechanical insulin pumps, or other means. We have already developed a sensor that shows promise. The sensor employs immobilized glucose oxidase coupled to a potentiostatic oxygen electrode. We have employed one version of the sensor as a central venous implant in dogs for over 100 days without the need for recalibration, and similar results from intravascular application in humans with another version have been reported by other groups. These successes were achieved by combinations of novel sensor design and the use of inherently stable enzymes. However, application in tissues, which is expected to be safer and more desirable, requires the development of new signal processing techniques. In Phase I, we developed novel signal processing algorithms suitable for sensors implanted in a complex tissue environment. In this Phase II project, the signal processing algorithms will be applied prospectively in extensive, new animal implant evaluations. This work represents significant innovation over previous work and is a crucial step toward implementation of a long-term implantable tissue-based glucose sensor.